em0 watchdog timeout with nfs

Jasper Berlijn freebsd-stable at claranet.nl
Thu Apr 5 21:11:47 UTC 2007


Jack Vogel wrote:
> On 4/5/07, Jasper Berlijn <freebsd-stable at claranet.nl> wrote:
>> Hi,
>>
>> At this moment I'm running FreeBSD 4.11 without any problems. A couple
>> of times I've tried to upgrade to 6.x but without any luck because of
>> the watchdog timeout errors on em0 when using nfs.
>>
>> Mar 30 11:30:48 large kernel: em0: watchdog timeout -- resetting
>> Mar 30 11:30:48 large kernel: em0: link state changed to DOWN
>> Mar 30 11:30:51 large kernel: em0: link state changed to UP
>> Mar 30 11:31:01 large kernel: em0: watchdog timeout -- resetting
>> Mar 30 11:31:01 large kernel: em0: link state changed to DOWN
>> Mar 30 11:31:03 large kernel: em0: link state changed to UP
>> Mar 30 11:31:20 large kernel: em0: watchdog timeout -- resetting
>> Mar 30 11:31:20 large kernel: em0: link state changed to DOWN
>> Mar 30 11:31:23 large kernel: em0: link state changed to UP
>>
>> em0 at pci0:14:0:  class=0x020000 card=0x002e8086 chip=0x100e8086 rev=0x02
>> hdr=0x00
>>     vendor   = 'Intel Corporation'
>>     device   = '82540EM Gigabit Ethernet Controller'
>>     class    = network
>>     subclass = ethernet
>>
>> When I try to copy a large file (1gb) then after a couple of mb's
>> (100mb) the wachtdog timeout will occur. The system is most of the time
>> idle. Using debug.mpsafenet="0" in /boot/loader.conf doesn't make any
>> difference.
>>
>> Any ideas?
> 
> The driver in 6.2 RELEASE fixed all known problems with
> watchdogs, other than REAL issues with the network/hardware.
> Have you tried installing that?
Running RELENG_6 (FreeBSD 6.2-STABLE (GENERIC) #0: Wed Mar 28 14:32:07
CEST 2007). The network is running at (1000baseTX <full-duplex>)

This system is dual boot so it's easy to switch from 4.11 (RELENG_4) and
RELENG_6. The system is running without any problem on 4.11.

- Jasper


More information about the freebsd-stable mailing list